The opportunity to express negative and defamatory statements prompted celebrities and ordinary people to abandon social platforms that are not equipped to deal with the problem.
Some publishers have chosen to close the comment sections. That’s why Google and Jigsaw (an incubator of parent Alphabet) are working on the project “Perspective,” which uses artificial intelligence to try to identify the “poisonous” comments, with the aim of reducing them.
The Perspective API released to developers recently returns a score that represents the comment is perceived as negative by users. This score can be used to develop functions such as automatic filtering of messages or to provide users with feedback on what they are writing before the actual publication.
Many online publishers have not wanted or have not been able to moderate comments to articles published on their sites effectively, and they chose to close them altogether, leaving open the discussion on social media.
For example, the New York Times, which has contributed to the development of Perspective, allows comments on about 10 percent of its products, due to the commitment required for moderation.
Google has a special interest in improving the situation. Its comment sections of YouTube can turn into a brawl, and comment moderation is left to online video creators. The Perspective API could simplify this task, in addition to an extensive list of responsibilities.
Interpretation of the comments from Perspective is imperfect, and Jigsaw recognizes the potential problems. The Perspective website offers users a way to test the comments and provide feedback on whether they are negative or not.
At this time, Perspective is available only in English, although some seem to understand the same comments in other languages.
Jigsaw will work with developers to improve Perspective developers and plans to release other Perspective API to add additional functions.